Lindenthal Recruited

Work on Davis' bridge was begun in the summer of 1880. It was to be a suspension bridge having two channel spans of 360 feet each and two shore spans of 180 feet each. Foundations for the channel piers were put in first and the piers were built up to an average height of ten feet each. Since the winter of 1880 was unusually severe, further work on the bridge was stopped. (28) None of the drawings for this abortive design seem to have survived.

In February 1881, the Bridge Company was reorganized and as a consequence all work on Davis' bridge was stopped and all prior contracts cancelled. The man who now held the controlling interest in the company's stock, David Hostetter, was also largely interested in Pittsburgh and Lake Erie and wished a different type of structure, because he thought it might be possible to run cars from his own line on the south bank to the lines of the Baltimore and Ohio on the north. Consequently a young German engineer, Gustav Lindenthal, was called in to make a design for a through-truss bridge. (29)

The new bridge engineer was also an immigrant. He was born in Brunn, Moravia, Austria-Hungary, and had been educated at the Provincial College of Brunn and at the polytechnical schools of Brunn and Vienna. He worked on railways in Austria and Switzerland before coming to America in 1874; in 1876 he assisted in the construction of buildings for the Centennial Exposition in Philadelphia. In 1881 he established himself in a private engineering practice in Pittsburgh. He was engaged in many railway and bridge projects, including the reconstruction of bridges on parts of what is now the Erie Railroad, various bridges in and near Pittsburgh, and railway surveys and estimates in Pennsylvania and neighboring states. When he reached the age of forty, he had established a reputation as one of America's great bridge engineers, and certainly the new Smithfield Street project was no small factor in his rise to fame.

In 1890 he set up a consulting office in New York City, devoting most of his time to bridge work. His best known works are the Queensboro Bridge (1901-1908) over the East River in New York, the Hell Gate Bridge (completed 1917) for the New York Connecting Railroad, and the Sciotoville Bridge (1914-1917) over the Ohio River. In 1902-03 he served as commissioner of bridges for the City of New York. In this capacity he advocated and established the practice of the association of engineers, in the design of large bridges, with architects whose special interest lay in the aesthetics of bridge construction.

As an engineer his greatest vision never materialized -- a bridge over the Hudson River at New York. From 1880 until his death he worked on the problem of transportation from New York to the Jersey shore and he constantly urged the adoption of his North River Bridge scheme. However, complications arising from decisions of the United States Army Engineers with reference to clearance, defeated final approval. The long span, 3,100 feet, heavy loading, and the huge costs of this project may be taken as a measure of Lindenthal's vision.

His designs were characterized by originality and boldness. He differed from many of his American contemporaries in his frequent choice of more complex structural forms and in some of his views as to working stresses. Like Roebling, he wrote much and he contributed to learned journals, many technical papers, chiefly on bridge design, but his chief monuments were his works. Pittsburgh is fortunate still to possess the first of his great designs which yet functions today, still serving its contiguous land areas and supporting weights that the engineer could not have foreseen when it was designed. (30)

Smithield Street Bridge (1881) Description

As in the case of Roebling, Lindenthal's is the best account of the construction of the Smithfield Street Bridge. (31) In 1881 "the writer was invited for consultation and to suggest changes in the plans, which should provide for a widening of the bridge by adding another roadway or track, should this ever become necessary in the future. After having submitting such plans, they were accepted and the writer was engaged to carry them out. This plan proposed to utilize the foundations and piers which had been commenced. They were to be built upon, without any offsets, to a width on top of 56 feet.

"As the width of the superstructure may ultimately reach 64 feet, or eight feet wider than the piers, it became necessary to let the sidewalks project over the masonry. The present width is 48 feet on the channel spans; the room for widening the bridge was left on the up-stream side. For the channel spans Pauli trusses (32) were proposed, 25 feet 8 inches apart, centre to centre, and centre line of the new floor (of 48 feet width) was shifted down-stream 8 feet 2 inches from the centre line of the old bridge. The sidewalk on the up-stream side was proposed to be detachable, so that the floor may be widened and the sidewalks again connected to it.

"For the approaches to the channel spans, plate girder deck spans, on lighter masonry piers, were proposed. This arrangement allows increasing the width of the bridge by simply adding more plate girders to each span on the piers which are long enough for that purpose. Being a deck bridge, it afforded an unobscured entrance view to the channel spans, the trusses of which were to rest on ornamental towers, giving to the superstructure an architectural appearance of strength and stability.

"The shifting of the centre line of new floor 8 feet 2 inches down-stream from the centre line of the old bridge allowed for erection of the new superstructure without stopping travel on the old bridge, in a manner described more in detail below.

"The Pauli truss type commended itself for the channel spans in this instance, for several reasons:

"1. The pleasing appearance (for a city bridge) in comparison with the ordinary parallel chord truss.

"2. The fact that the trusses could be made high in the middle (without detriment to their stability in case of high winds), thereby reducing the chord strains and chord sections. In connection with the light and slender web-members, it permitted the economy in the trusses of over 9 percent, as compared with parallel chord trusses (with inclined end posts) of same height (50 feet). The deflection and vibration of high trusses is small and their rigidity is great.

"3. The bottom chord or cable is exposed to the sun's ray as much as any other truss member; therefore unequal temperature effects in the trusses are avoided. The covered floor construction is independent of the trusses as to temperature effects.

"4. The floor had to be cambered 18 inches in each 360 foot span to agree with the general grade of the new bridge. A straight bottom chord with a rise of 18 inches in 360 feet was undesirable.

"At first it was proposed to build the new structure 15 feet higher at highest point than the old bridge. But the river men, in the interests of navigation demanded the structure be at least 20 feet higher, or 57 above low water mark, to which the Bridge Company objected, on the ground that the additional 5 feet would injure travel over the bridge much more, by reason of a steep grade at the Pittsburgh end, than it would benefit navigation.

"There is no statute prescribing the height of bridges over the Monongahela River. The case was taken to a court of equity, and argued there by lawyers pro and contra, resulting in a preliminary injunction against the Bridge Company building the bridge lower than 20 feet. To continue the litigation would have required much time. After a suspension of work at the bridge for 10 months the Bridge Company decided to accede to the demands of the river men.

"The following is a description of the material and methods used in the construction of the bridge:

Masonry

The masonry consists of a gray, hard and durable sandstone, free from admixtures of clay or iron oxide particles. It was quarried near Homewood, Pa. on the Pittsburgh and Lake Erie Railroad where it is found in large blocks of 100 to 500 cubic yards, without any stripping. The masonry is rock-faced, with drafts 1 inch wide all around the face of the stones, which are in courses of alternate headers and stretchers.

"The dimensions of the stones are 24 inches to 16 inches in thickness, 7 feet to 4 feet in length, 3 feet to 11 feet in width, with beds and joints dressed regular and true. The backing for the abutments and wing walls consisted of regular shaped stones, with dressed beds; for the heart of the piers concrete filling was used. It was applied in layers of 12 inches thick. It proved superior in every way to ordinary stone backing. Iron clamps bind the stone in the pier heads in every course.

"The use of spalls was not permitted in any part of the masonry. All spaces between stones were filled with concrete, rammed with iron rammers, making every course absolutely water-tight. Great attention was given to the bond. The stone blocks were laid in alternate header and stretcher courses, which made the coincidence of stone joints in the heart of the pier impossible. In this way each stone is bonded in every direction. The concrete backing, after setting, was very hard and tough; it adhered to the stones with great tenacity, and made the piers monolithic in fact.

"In the execution of the work care was taken to set every stone immediately before setting. When laid in position the stone was settled by repeated blows of a heavy wooden ram. Any stone breaking under this operation was removed.

"The face joints of the finished masonry were cleaned out to a depth of 1 inch, and thoroughly moistened, and caulked with Portland cement and sand mortar, mixed one to one.

"For all face masonry exposed to the weather American Portland cement was used for the mortar; for concrete backing and foundations, Rosendale cement was ordinarily used.

"All cements were required to be so finely ground that 90 percent of the whole would pass through a sieve of 50 meshes to the lineal inch. Tests as to its tensile strength were conducted on a Fairbanks testing machine with moulded briquettes of pure cement.

"Rosendale cement made of a stiff paste, having been one day in water and one day in the air, at an even average temperature of 70 degrees Fahrenheit (in a room), were tested to show the tensile strength of at least 40 pounds per square inch.

"American Portland cement briquettes, under the same conditions, were tested to show a tensile strength of at least 80 pounds per square inch.

"Similar briquettes, after having been four days in water and one day in the air, at the above average temperature, were tested for a tensile strength of 60 pound per square inch for Rosendale cements, and 150 pounds for American Portland cement.

"The concrete used throughout the work was composed of 2 parts of sound broken stone, passing through a 3 inch ring; 2 parts of clean gravel from the size of a pea to 2 inches diameter; 2 parts of washed river sand; 1 part of Rosendale cement of accepted quality.

"For concrete under water 2 parts of cement were used to allow for waste by washing in depositing it under water. With a little care in the operation the loss, however, was insignificant. The stone, gravel and sand were mixed on a board platform, then cement added, and the whole mass thoroughly rehandled in a dry state. Water was then added in barely sufficient quantity to reduce the whole mass, by lively and severe shoveling, to a stiff mortar. This was put immediately in place in layers of not over 12 inches thick, and thoroughly rammed with iron rammers about 5 inches square and weighing 36 pounds, until mass flushed uniformly over the whole surface.

"For depositing concrete under water for the pier foundations square wooden troughs were used, reaching down almost to the bottom, and the concrete dumped in and raked even with iron rakes having long handles. The running out of the concrete was prevented by sheet piling. When a change in the masonry of Pier No. 4 required the removal of a few stones they were found to form with the concrete backing one solid mass, which had to be rent asunder with steel wedges and sledge-hammer, and would sometimes break through the stone rather than through the concrete.

"Openings or slots for one car track were left in the new abutments and piers to accommodate travel on the old bridge.

"The pier posts of the channel spans on the down-stream side have their bearing near to the pier ends, and to prevent cracking of the channel piers or uneven settlement after the superstructure should be in place, riveted iron anchors were walled into the top of piers Nos. 2, 3 and 4.

"The coping on the piers, consisting of two projecting courses of cut stone, was nearly all in place for a grade 15 feet higher than the old bridge, at the time of the dispute with the river men. When the height of the piers was increased to suit a grade 20 feet higher than the old bridge, the additional masonry was built on top of the coping in the form of pedestals of cut stone.

"After the erection of the superstructure had so far progresses that travel could be turned on to one track on the new bridge, the old bridge was abandoned, and the taps and openings in the masonry of the new abutments and piers successively walled in and closed. In this way it was possible to complete the masonry work without stopping travel on the old or new bridge.

Superstructure

"The roadway is at present 22 feet 10 inches wide in the clear, and two sidewalks each 10 feet in the clear. The full width of the bridge on the deck spans of approaches is 43 feet 6 inches, and on the channel spans, which are through spans, 48 feet.

"The bridge can be widened out, if ever required, to 64 feet. This made it necessary to erect the present superstructure nearer to the down-stream end of the piers. It detracts much from the appearance of the bridge, which is unsymmetrical at present.

"It was important not to stop travel during the rebuilding of the bridge. Passengers and freights from and to the Pittsburgh and Lake Erie Railroad must pass over it. Besides, there is a heavy traffic in coke, iron and other mill material, which would have been compelled to take a long, roundabout way. The construction of the superstructure had to be arranged to allow of the erection first of one track and then of the other.

"If the new bridge had really been built 15 feet instead of 20 feet higher than the old one there would not have been left height enough near the ends of the channel spans for teams to pass under on the old bridge. It was therefore intended to erect the channel spans about 5 feet higher than their proper grade, and to complete the floor and tracks of the same.

"The pier posts would have temporarily rested on sand jacks, by means of which both spans, weighing about 1600 tons when completed, could have been simultaneously lowered in a few hours to their proper grade. One track and sidewalk on the plate girder approaches on the down-stream side would have been meanwhile prepared for use. In this way travel would have been interrupted only for one day. But this operation became unnecessary when the new grade was raised 20 feet above the old bridge.

Channel Spans

"It was found that the use of steel, in the trusses at least, would prove economical as compared with wrought iron. The saving based on the prices at that time was over $21,600.

"The Pauli trusses were designed with an uneven number of panels, namely 13, in order to get two tangential points of attachment for each truss to the floor-construction, thereby securing greater longitudinal and transverse rigidity of the entire bridge frame. Roller bearings for pier posts were avoided; the middle posts, supporting two truss ends each, have a fixed and square bearing on heavy pedestal castings on the pier. Each end post has a bearing on a 6 inch steel pin in a cast-iron pedestal on which it can rock. It is probable that very little movement takes place on account of friction on the pin, and that the posts would bend or spring. The resulting bending moments on the end posts have been considered in proportioning them.

"The projected length, 27 feet 7-5/8 inches, of all panels being alike, it follows that the lengths of chords in a curved line are unlike, and if the curve were a circle or a parabola, then the angles formed by the straight chord sections would also all be unlike.

"For practical reasons it is desirable to have these angles all alike, so as to have only one template for the beveled joints. This condition would prescribe the character of the curve, in this instance a sine-curve. The difference in curvature between a sine-curve and arc of a circle was found to be small (2-1/2 inches). The difference in the bevel points was inappreciable (3/64 inch). Therefore a true arc line was then assumed for the chords to facilitate other calculations.

"The vertical web-members are in tension from the dead load or from a uniformly distributed live load. They will sustain compression strains only from an uneven distributed load. Near the centre of truss they are long and slender, requiring intermediate bracing, which was placed at half the truss height for the entire length of trusses.

"The suspenders from trusses to floor, which were all stiffened to prevent vibration, were not made adjustable; their exact lengths were calculated to give the required camber of 18 inches to the floor construction. The truss camber was obtained by shortening the lower and lengthening the upper chord members 3/16 inch, so that after erection it amounted to 2 inches.

"All diagonal bars were made adjustable and single; they are strained from partial loads only. The trusses were adjusted to their proper shape by means of these ties, which received a slight initial strain.

"The top and bottom chords, pier-posts, diagonal-ties, and pins are of steel; all other parts are of wrought-iron with steel rivets. The calculated sections of the vertical web-members for steel were so light that for practical reasons they were all made of wrought-iron and of the same section.

Quality of Steel Used

"Every heat of steel was tested and its quality determined before any more work was done to it.

"For the compression members and pins, the steel was required to stand the following test on specimen bars 3/8 inch diameter:

Elastic limit: 50 to 55,000 pounds per square inch.

Elongation in 8 inches: Minimum 12 per cent.

Reduction of area at fracture: Minimum 20 per cent.

Cold bending: 180 degrees around its own diameter without crack.

Cold punching of holes in flat 3 x 1/4 inch bars; 3/16 inch from the edge without crack or distention of metal.

"All specimens and shapes were required to be finished at nearly the same heat, as it was observed that rods finished at a lower heat would give higher tension results than samples of same steel finished at a higher heat.

"The Andrew Kloman firm in Pittsburgh had contracted to procure the steel and to furnish the steel shapes.

"The intention was to use Bessemer steel for the compression members; a large lot of Bessemer steel was tested, but few samples were found to stand the required tests. The difficulty seemed to consist in controlling the uniformity of the steel within close limits for quality and strength. After a while the attempt was given up and open hearth steel was substituted. No trouble was then experienced in getting a uniform grade of steel of prescribed quality.

"The top chord sections consist of four leaves, which were originally designed to be each a 20 inch steel plate with 4 x 4 inch singles for flanges. In ordering the steel it was discovered that enough plates of that width could not be procured in the required time. Therefore, the chord sections were changed to 10 inches and 12 inches steel plates, with 4 x 4 inch angles, composed as shown in the drawings.

"Notwithstanding the great care used, the finished plates and angles were by no means a uniform product. According as they in rolling were finished at a higher or lower heat, they would have different degrees of hardness. Steel plates and angles finished at a lower heat had a smooth surface, and the noise of punching them resembled pistol-shots, while plates finished at a higher heat had a rougher surface, and there was hardly more resistance to punching than in wrought-iron.

"The specifications for riveted steel work provided that the punched rivet-holes, 3/4 inch diameter, should in the assembled parts be enlarged to 1 inch diameter by reaming. The time for the delivery of the steel work growing short, the question was considered whether the reaming of the holes could be avoided, to hasten the completion of the work at the shops. Messers. Kellogg and Maurice, in Athens, Pa., had the contract for this part of the work.

"To that end and the following experiments were made:

"Ten specimens were cut from the same steel plate 1/4 inch thick; one specimen was tested to ascertain the tensile strength of the steel in the specimen. The nine other specimens, all alike in form, were prepared as shown in sketch, for the purpose of ascertaining the effects of punching holes, of punching and reaming, and of drilling. The tests were expected to show the amount of reaming required, and whether any annealing effects from the hot rivet on the injured steel around the punched hole could be observed.

"The conclusion from these tests was that the injured steel (of the quality used in this instance) around the punched hole was in part restored by annealing in contact with the hot rivets, the size of which was large in proportion with thickness of steel plates and angles as used in the chords.

"The reaming of the punched holes to a greater extent than to make the rivet holes smooth and straight was therefore dispensed with, and a reduction in the price for the finished work agreed upon.

"The same quality of steel as for the compression members was used for them; they were forged from solid steel billets, and turned to size. No appreciable difference in the hardness of the metal in the pins was observed.

"For tension members and rivets, the steel was required to stand the following tests on specimen bars 5/8 inch diameter:

Elastic limit: 45 to 40,000 pounds per square inch -- Yield.

Ultimate strength: 70 to 80,000 pounds per square inch.

Elongation in 3 inches: Minimum 18 per cent.

Reduction of area at fracture: Minimum 30 per cent

Cold bending: to a loop 360 degrees around its own diameter, without crack.

Cold punching in 3 x 3/4 inch bars of 1 inch rivet holes: 1/4 inch from the edge without crack or distension of metal.

Open hearth steel of the above and uniform quality was obtained without trouble.

"The eye-bars were made by the Kloman process, i.e., the bars were rolled from billets between reversible and adjustable rolls, in such manner as to leave the ends thicker than the bar. The ends were then spread and forged to the proper shape of the eye, under a steam hammer. The heaviest steel bars for this bridge were 28 feet 6-1/2 inches long, centre to centre of eyes, and 1-13/16 inches thick. All steel billets and all steel bars required very close inspection for flaws, the detection of which was sometimes difficult.

"It has been stated that for the detection of flaws in steel or iron, a magnetic needle had been used with success, though the manner of its use the writer has not heard stated. A device for the certain discovery of flaws in steel bars is certainly needed. Where the solid metal sections are proportioned very economically to the work they have to do, flaws are a source of great danger, especially in attenuated steel structures; flaws in wrought-iron are more likely to happen in the direction of the fibre, but in steel they can as well happen cross-wise to the direction of the tension strain as any other way.

"Three steel bars 9 feet long between centres of eyes, and 4 inches x 1-1/16 inches in section were tested to ascertain the effect, if any, of annealing the finished bars. The results were as follows:

Plate Girder Spans

"There are six plate girders in each span beneath the flooring, namely, one girder under each rail and one girder under each sidewalk, which is detachable on the up-stream side.

"This arrangement was chosen to admit of the erection first of the new down-stream street track, which came to lie sideways of, but on a higher grade than, the down-stream track of the old bridge. To this track travel was confined during erection. Plate girders were chosen for this reason that for the limited depth of floor (for a grade 15 feet higher than old bridge, as at first contemplated), it gave a more rigid construction than open girders of low depth. It was also more convenient to work into them, and get rid of a lot of wrought-iron which was on hand, and was left over from orders for the suspension bridge originally intended to be built.

"Could the writer have foreseen that the new grade would be 20 feet higher than the old bridge, the deck spans of the approaches would have been made of two open girders of greater depth, in a manner that would have admitted of finishing both tracks on them at the same time. This would have been also more economical. As it was, the plate-girders were nearly finished when the change in grade was made.

"For all wrought-iron work in the bridge the quality of iron was required to be equal to that of standard bridge iron.

"Steel rivets were used for all wrought-iron bridge-members and girders.

Previous Page   --   Page 3 of 4   --   Next Page

Page created:
Last modified: 30-Sep-1999

Historic American Engineeering Record (HAER) Text: James D. Van Trump, 1974